John Haines USGS Coastal (Climate) Change Activities - A Foundation in Observations 86th Coastal Engineering Research Board Meeting San Diego, CA 3 June 2009 John Haines USGS, Coastal and Marine Program Coordinator

Issue: Coastal Change - from Storms to Sea-Level Rise (modified after Bindoff, 2007; Rahmstorf, 2007) Short-term Variance (hours to decade) Storm impact/recovery Annual cycles El Niño Long-term Trend (decades to centuries) Sediment deficit or surplus Sea-level rise Focus – How Risk and Vulnerability evolve in response to natural and human factors.

Complex Systems + Complex Responses  Comprehensive, Integrated Research Multiple human and natural drivers spanning multiple time scales Diversity of systems – glaciated coasts to tropical atolls, wetlands, and barriers responding dynamically Observations – Research – Modeling Needs span policy/management scales – National and Regional Modeling/Assessment needs from simple to complex

A Schematic of the Process Required Input LIDAR OBS Evaluate output Area for collaboration: prioritize national observation resources to provide accurate and up-to-date elevation data Area for collaboration: prioritize national observation resources to minimize uncertainty LIDAR OBS Bathy/Topo low Overwash and Erosion models + Infrastructure Risk Habitat Risk Low High Med. Weather high medium Observations Processes Response Probability Risk Analysis Areas for collaboration 1. Nested modeling using national observation resources and large scale models to support high resolution models—we need accurate Boundary Condition inputs 2. Scenario exploration for likely climate changes and extreme storms wave/water OBS wave/water MODELS

Partnering: USACE and USGS Observations – National (Lidar) shoreline characterization, Storm response, field test beds (FRF) Regional characterization and modeling (Fire Island, SW Washington, Gulf Coast, Carolinas) Model Development – Commercial, Research, Applied Distribution and Volume of Holocene Sediment on Inner Shelf Relates to Migration Rate of Barrier-Island System

National Observations, Research & Products

National Observations, Research & Products Storm Vulnerability Assessment Coastal Change Assessment Surge Monitoring Sea-level Rise Vulnerability Index

Regional Observations, Research & Products - Focus on Geologic Setting & Processes, Sediment Inventory and Budget A A’ A A’

Cape Fear Cape Romain Regional Observations, Research & Products - Modeling Sediment Transport and Coastal Evolution Cold Front net flux to the NE Tropical Storm net flux to the SW

Regional Observations, Research & Products Chandeleur Islands XBEACH simulations Sediment Volumes Subaerial Change STWAVE/ADCIRC

Moving Forward: Science for Decision-making in response to Sea-Level Rise Explicitly including uncertainty Explicitly including management application Extracting information from data/information resources

Coastal Vulnerability Index Input Data: Coastal Vulnerability Index (Thieler and Hammar-Klose, 1999) Utilized existing data for six geological and physical process variables (~ 8km grid): Geomorphology Historic shoreline change Coastal slope Relative sea-level rise rate Mean sig. wave height Mean tidal range Solve this differential equation? d(state)/dt = funct.(geomorphology, wave-climate, sea level, etc.) OR, solve this probabilistic version P(state | inputs) = Bayes Rule

Bayesian method: Predict SLR impact on coastal erosion Data (and uncertainty) are input Prediction of probability of all possible outcomes is output Straightforward to evaluate likelihood of outcome to exceed a user-specified tolerance prob. erosion>2m/yr =28.95% inputs SLR tide wave geormorph

Evaluate probability of erosion given different SLR ranges 1. Select SLR category Increasing Sea Level 2. Examine P (SLC < -1 m/yr) For Very High SLR: P(SLC < -1 m/yr) = 56.9%

Mapping Erosion Probabilities Probability of Erosion > 2 m/yr Mapping Erosion Probabilities Boston New York D.C. Coastal data sets can be evaluated with Bayesian network. Map probability of critical scenarios. Atlantic Ocean Charleston Miami

Prediction Uncertainty Certainty of most likely outcome (probability) Prediction Uncertainty high confidence We can also map uncertainty which can be used to identify where we need better information. Areas of low confidence require: better input data better understanding of processes Use this map to focus research resources low confidence high confidence

Observational Requirements Data/Observations integrate for us; capturing processes we don’t fully understand and including information on uncertainty Data collection should be designed to lead to understanding of processes of interest/importance, reduce uncertainty in prediction, and inform improved data collection This means more data – but it also means more thoughtful “parameterization” of how we describe the system, the forcing, and the response; and Observations must be consistent & span scales appropriate for a variety of decision-making needs.